Source
- URL: https://x.com/stuartcarnie/status/1947566650276123023
- Author: Stuart Carnie (@stuartcarnie)
- Posted: 2025-07-22 07:57:28
Branch
1/ @stuartcarnie
@nicbarkeragain
It also matters how random the data is. If the branch predictor can identify patterns, the results will be much closer. For example setting the data to i % 3 results in <3x difference
2/ @nicbarkeragain
@stuartcarnie
Yes, the gap will close significantly if the branch predictor is able to do a good job with the data at hand!